Origin of Rashba-splitting in the quantized subbands at Bi2Se3 surface

TL;DR

This study investigates the origin of Rashba spin-splitting in quantized subbands at the Bi2Se3 surface, revealing that differences in spin-orbit coupling, rather than surface potential gradients, cause the observed splitting behavior.

Contribution

It demonstrates that the Rashba-splitting discrepancy between conduction and valence band subbands is due to differences in spin-orbit coupling strength, supported by experimental ARPES data and first-principles calculations.

Findings

CB subbands are strongly Rashba spin-split

VB subbands do not show clear spin-splitting

Differences in spin-orbit coupling explain the splitting discrepancy

Abstract

We study the band structure of the $\text{Bi}_2\text{Se}_3$ topological insulator (111) surface using angle-resolved photoemission spectroscopy. We examine the situation where two sets of quantized subbands exhibiting different Rashba spin-splitting are created via bending of the conduction (CB) and the valence (VB) bands at the surface. While the CB subbands are strongly Rashba spin-split, the VB subbands do not exhibit clear spin-splitting. We find that CB and VB experience similar band bending magnitudes, which means, a spin-splitting discrepancy due to different surface potential gradients can be excluded. On the other hand, by comparing the experimental band structure to first principles LMTO band structure calculations, we find that the strongly spin-orbit coupled Bi 6$p$ orbitals dominate the orbital character of CB, whereas their admixture to VB is rather small. The spin-splitting discrepancy is, therefore, traced back to the difference in spin-orbit coupling between CB and VB in the respective subbands' regions.